38
Eigenvalue Eigenvector Problem

Applied numerical methods lec13

Embed Size (px)

Citation preview

Page 1: Applied numerical methods lec13

Eigenvalue – Eigenvector Problem

Page 2: Applied numerical methods lec13

EIGENVALUES & EIGENVECTORS

The eigenvalue problem is a problem of considerable

theoretical interest and wide-ranging application.

For example, this problem is crucial in solving systems of

differential equations, analyzing population growth models,

and calculating powers of matrices (in order to define the

exponential matrix (A100)).

Other areas such as physics, sociology, biology, economics

and statistics have focused considerable attention on

“eigenvalues” and “eigenvectors”-their applications and their

computations

Page 3: Applied numerical methods lec13
Page 4: Applied numerical methods lec13

Eigenvalue Problems(Mathematical Background)

0xaxaxa

0xaxaxa

0xaxa xλa

nnn2n21n1

n1n222121

n1n212111

0XA

The roots of polynomial D(λ) are the eigenvalues of the eigen system

A solution {X} to [A]{X} = λ{X} is an eigen vector

(homogeneous system)

0 XIA

IA tDeterminanD

(eigen system)

Page 5: Applied numerical methods lec13
Page 6: Applied numerical methods lec13

Example 1

Page 7: Applied numerical methods lec13
Page 8: Applied numerical methods lec13

CW

Page 9: Applied numerical methods lec13

Example 2

Page 10: Applied numerical methods lec13
Page 11: Applied numerical methods lec13
Page 12: Applied numerical methods lec13
Page 13: Applied numerical methods lec13
Page 14: Applied numerical methods lec13
Page 15: Applied numerical methods lec13

15

The Power Method(An iterative approach for determining the largest eigenvalue)

Example (3):

Iteration 1: initialization [x1, x2, x3]T = [1 1 1]T

Iteration 2: A [1 0 1]T

Iteration 3: A [1 -1 1]T

Iteration 4: A [-0.75 1 -0.75]T

Iteration 4: A [-0.714 1 -0.714]T

(Exact solution = 6.070)

Page 16: Applied numerical methods lec13

Class Work: Use the power method to find the dominant eigenvalue and

eigenvector for the matrix

Page 17: Applied numerical methods lec13
Page 18: Applied numerical methods lec13
Page 19: Applied numerical methods lec13

Example 4

Page 20: Applied numerical methods lec13
Page 21: Applied numerical methods lec13

21

Power Method for Lowest Eigenvalue(An iterative approach for determining the lowest eigenvalue)

422028101410

281056202810

1410281042201

...

...

...

A

7510

1

7510

1241

8840

1241

8840

1

1

1

422028101410

281056202810

141028104220

.

.

.

.

.

.

...

...

...

Example (5):

Iteration 1:

Iteration 3:

Exact solution is 0.955 which is the reciprocal of the smallest eigenvalue,

1.0472 of [A].

Iteration 2:

7150

1

7150

9840

7040

9840

7040

7510

1

7510

422028101410

281056202810

141028104220

.

.

.

.

.

.

.

.

...

...

...

7090

1

7090

9640

6840

9640

6840

7150

1

7150

422028101410

281056202810

141028104220

.

.

.

.

.

.

.

.

...

...

...

Idea: The largest eigenvalue of [A]-1 is the reciprocal of the lowest

eigenvalue of [A]

5563

422077810

778155632810

077815563

7781 .

..

...

..

.A

Page 22: Applied numerical methods lec13

Faddeev-Leverrier Method for Eigenvalues

Page 23: Applied numerical methods lec13
Page 24: Applied numerical methods lec13
Page 25: Applied numerical methods lec13
Page 26: Applied numerical methods lec13
Page 27: Applied numerical methods lec13

Use Faddeev’s Method to find the eigenvalues of the following matrix

Page 28: Applied numerical methods lec13

Eigenvalue Problems(Physical Background 1)

1212

1

2

1 xxkkxdt

xdm

tsinAx ii

Mass-spring system

Analytical solution

(vibration theory)

2122

22

2 kxxxkdt

xdm

k kAi = the amplitude of the vibration of mass i

and ω = the frequency of the vibration, which is

equal to ω = 2π/Tp, where Tp is the period.

1

2

3

tsin-Ax 2ii 4

To Find A1, A2, and :

Substitute equations 3 and 4 into 1 and 2

0A A

0A A

21

21

2

22

1

2

1

2

2

m

k

m

k

m

k

m

k

The eigenvalues to this system are the

correct frequency , and the eigenvectors

are the correct A1 and A2.

take 2 as λ

Page 29: Applied numerical methods lec13

0A A

0A A

21

21

2

22

1

2

1

2

2

m

k

m

k

m

k

m

k

The eigenvalues to this system are the

correct frequency , and the

eigenvectors are the correct A1 and A2.

take 2 as λ

0xaxaxa

0xaxaxa

0xaxa xλa

nnn2n21n1

n1n222121

n1n212111

0XA (homogeneous system)

0 XIA (eigen system)

Page 30: Applied numerical methods lec13

30

Eigenvalue Problems(An instance of mass-spring problem)

0A A

0A A

21

21

2

22

1

2

1

2

2

m

k

m

k

m

k

m

k

mode second s

mode first s-

-

1

1

2362

8733

.

.

k=200 N/m

First mode: A1 = -A2

Second mode: A1 = A2

0A A

0A A

21

21

2

2

105

510

m1=m2=40 kg

Page 31: Applied numerical methods lec13

A unique set of values cannot be obtained

for the unknowns. However, their ratios can

be specified by substituting the eigenvalues

back into the equations. For example, for

the first mode (ω2 = 15 s−2), Al=−A2. For the

second mode (ω2 = 5 s−2), A1 = A2.

Note:

Page 32: Applied numerical methods lec13

Eigenvalue Problems(Physical Background 2)

Page 33: Applied numerical methods lec13
Page 34: Applied numerical methods lec13
Page 35: Applied numerical methods lec13

Eigenvalue Problems(Physical Background 3)

Page 36: Applied numerical methods lec13
Page 37: Applied numerical methods lec13

Special Problem 5

Page 38: Applied numerical methods lec13

Home Work

1.

2.

3.